Insights into the Binding of Pyridines to the Iron−Sulfur Enzyme IspHIngrid Span,*,†,‡ Ke Wang,§ Wolfgang Eisenreich,† Adelbert Bacher,† Yong Zhang,∥ Eric Oldfield,§

and Michael Groll†

†Center for Integrated Protein Science Munich, Chemistry Department, Technische Universitat Munchen, Lichtenbergstrasse 4,85747 Garching, Germany§Department of Chemistry, 600 South Mathews Avenue, University of Illinois at Urbana−Champaign, Urbana, Illinois 61801, UnitedStates∥Department of Chemistry, Chemical Biology, and Biomedical Engineering, Stevens Institute of Technology, Castle Point onHudson, Hoboken, New Jersey 07030, United States

*S Supporting Information

ABSTRACT: (E)-1-Hydroxy-2-methylbut-2-enyl 4-diphos-phate reductase (IspH) is a [Fe4S4] cluster-containing enzymeinvolved in isoprenoid biosynthesis in many bacteria as well asin malaria parasites and is an important drug target. Severalinhibitors including amino and thiol substrate analogues, aswell as acetylene and pyridine diphosphates, have beenreported. Here, we investigate the mode of binding of fourpyridine diphosphates to Escherichia coli IspH by using X-raycrystallography. In three cases, one of the iron atoms in thecluster is absent, but in the structure with (pyridin-3-yl)methyldiphosphate, the most potent pyridine-analogue inhibitor reported previously, the fourth iron of the [Fe4S4] cluster is presentand interacts with the pyridine ring of the ligand. Based on the results of quantum chemical calculations together with thecrystallographic results we propose a side-on η2 coordination of the nitrogen and the carbon in the 2-position of the pyridine ringto the unique fourth iron in the cluster, which is in the reduced state. The X-ray structure enables excellent predictions usingdensity functional theory of the 14N hyperfine coupling and quadrupole coupling constants reported previously using HYSCOREspectroscopy, as well as providing a further example of the ability of such [Fe4S4]-containing proteins to form organometalliccomplexes.

■ INTRODUCTION

Isoprenoids, including steroids and terpenes, constitute one ofthe largest and most diverse class of natural products. In allorganisms they derive from the isoprene derivatives isopentenyldiphosphate (IPP, 1) and dimethylallyl diphosphate (DMAPP,2).1 Two different biosynthetic pathways are known to produceboth IPP and DMAPP: the mevalonate pathway present inmammals as well as some microorganisms, and the 1-deoxy-D-xylulose-5-phosphate (DXP) pathway found in most patho-genic bacteria including Mycobacterium tuberculosis, as well asthe malaria parasite, Plasmodium falciparum.2,3 Due to use ofthe mevalonate pathway in humans, enzymes of the DXPpathway are important anti-infective drug targets.The final step in the DXP pathway is the conversion of (E)-

1-hydroxy-2-methylbut-2-enyl 4-diphosphate (HMBPP, 3) intoa ∼6:1 mixture of IPP and DMAPP (Scheme 1).4 Thisreductive dehydroxylation reaction is catalyzed by HMBPPreductase (IspH), an oxygen-sensitive iron−sulfur enzyme.5,6

Despite the identification of several intermediate species of thisreaction, the reaction mechanism is still the topic of livelydebate.4,7−14

IspH has a trefoil-like architecture with a [Fe4S4] cluster inthe active site (Figure 1), serving both structural and catalytic

roles.15,16 The decomposition of this oxygen-sensitive cofactorleads to partial loss of the tertiary structure and to completeloss of function,12 as observed also for aconitase17 and radical S-adenosyl-L-methionine (SAM)18,19 enzymes. Catalysis takesplace at the unique iron site that is not coordinated by anycysteine side chain, unlike the three remaining iron sites.9 TheHMBPP substrate binds directly to the unique fourth iron viathe hydroxyl oxygen atom in the early stage of catalysis (FigureS1a in the Supporting Information).9 Recently, it was shownthat several oxygen, nitrogen, and sulfur atom-containingsubstrate analogues are also able to coordinate to this uniqueiron site (Figure S1b,c in the Supporting Information). These

Received: February 4, 2014

Scheme 1. Reaction Catalyzed by IspH

Article

pubs.acs.org/JACS

© XXXX American Chemical Society A dx.doi.org/10.1021/ja501127j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

observations are of great interest in the context of inhibitordevelopment.20,21

Since its key role in the biosynthesis of isoprenoids wasdiscovered, IspH has become the subject of intensive research,particularly in regard to the development of new antimicrobialagents.22 Several studies have identified a variety of compoundsthat bind to the active site of IspH and inhibit its activity.10,23,24

Most of these molecules have a diphosphate group that binds ina similar way to IspH as does the HMBPP substrate. Structural,spectroscopic, and computational studies25 of IspH interactingwith derivatives of the substrate 3 (Scheme 2), in which the

hydroxyl group is replaced by an amino (4) or thiol (5) group,have shown that the heteroatoms coordinate to the unique ironsite.20 Furthermore, crystallographic studies have revealed thepromiscuous reactivity of IspH, hydrating acetylenes 6 and 7 tothe aldehyde 8 and the ketone 9, respectively,21 with theenolate of 8 binding to the fourth iron and stabilizing theprotein with respect to cluster decomposition in the presenceof atmospheric oxygen. The importance of IspH as a new drugtarget and its versatile and flexible catalytic site thus providemotivation for the characterization of inhibitors that may benew drug leads.In addition to the linear compounds that, structurally, are

closely related to 3, the pyridine derivatives 10−13 (Scheme 2)have also been shown to inhibit IspH enzymatic activity.23

Moreover, electron paramagnetic resonance (EPR) as well asX-band hyperfine sublevel correlation (HYSCORE) spectro-scopic studies have indicated that 10 interacts with the uniqueiron of the [Fe4S4] cluster in the active site of IspH.

26 What has,however, been unclear is just how the pyridine inhibitors bindinto the active site.In early work we used computational docking to propose that

the pyridine inhibitors bound to reduced IspH as illustrated inFigure 2a. The aromatic ring in the inhibitor is located close tothe fourth iron, but we speculated that most likely a Coulombicinteraction between the pyridinium ring and the E126 carboxylwas important for ligand binding.23 In later studies we usedHYSCORE spectroscopy (Figure 2b) to investigate the bindingof 10 to 15N-labeled IspH finding that there was a large 14Nhyperfine coupling (∼7 MHz) and that the nuclear quadrupolecoupling constant (NQCC) was ∼3 MHz. These values aresimilar to those found for aromatic bases bound to Fe in bothproteins as well as model systems,26 leading to the idea that 10might bind end-on in the reduced protein, as shown in Figure2c.Here, we have investigated the binding of 10, as well as

several analogues, to IspH by using X-ray crystallography. Wealso used density functional theory (DFT) methods to probethe nature of the bonding between the ligand and metal cluster,in addition to computing the HYSCORE observables: thehyperfine and NQCC values.

■ RESULTS AND DISCUSSION

Crystallographic Structures. To investigate the mode ofbinding and inhibition we crystallized (pyridin-3-yl)methyldiphosphate (10) with Escherichia coli IspH and determined the

Figure 1. Structure of E. coli IspH. (a) Surface representation of themonomeric IspH with the surface in transparent gray and the cartoonmodel colored according to secondary structure elements (α-helices inblue, β-sheets in yellow and loops in gray). (b) The [Fe4S4] cluster inthe active site is shown as a ball-and-stick model, with the iron atomscolored in orange and the sulfur atoms in gold. The cluster binds tothe protein via three cysteine residues; other ligands can bind to thefourth iron site.

Scheme 2. Structures of Compounds That Interact with IspH

Figure 2. Predicted models and a 9 GHz 14N/15N HYSCORE spectrum for pyridine inhibitors binding to IspH. (a) Docking pose. Reprinted withpermission from ref 23. Copyright 2010 American Chemical Society. (b) 9 GHz HYSCORE result for [14N] 10 binding to [15N]-labeled IspH.Reprinted with permission from ref 26. Copyright 2011 American Chemical Society. (c) Proposed end-on binding of pyridine to a [Fe4S4] cluster.Reprinted with permission from ref 26. Copyright 2011 American Chemical Society.

Journal of the American Chemical Society Article

dx.doi.org/10.1021/ja501127j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXXB

structure to 1.7 Å resolution (PDB ID 4MUX); for details seeTable S1 in the Supporting Information. The overallarchitecture of IspH in this complex is virtually identical tothe previously reported structure of the protein bound to thesubstrate 3.9 The pyridine ring is located close to the apical ironof the cluster, however, this labile iron site is only present in theIspH−10 complex (Figure 3). Upon inspection of the apicaliron, the difference electron density displays a negative peak atthis site indicating that this site is partially occupied and, aspreviously noted, the unique fourth iron of the cluster readilydissociates when not stabilized by a ligand.12,21 Comprehensiveinvestigation (Table S2 and Figure S2 in the SupportingInformation) of the occupancy of the fourth iron (Fe3 in thePDB file) resulted in values of 0.45 for chain A and 0.51 forchain B indicating that there is a [Fe4S4] cluster in about 50%of the molecules in the IspH−10 crystal. The occupancy of theligand, on the other hand, is 1. The position of the ligand 1o isvery similar to that predicted earlier using computationaldocking23 with a root-mean-square deviation (rmsd) of 0.89 Åbetween the ligand’s heavy atoms in the docked and crystalstructures, shown in Figure 3c.The crystal structure of 10 bound to IspH reveals two main

points of interest. First, there are two pyridine ring atoms thatare very close (2.3−2.4 Å) to the fourth iron. Second, there isevidence for a continuous electron density between the fourthiron and the pyridine nitrogen. Taken together, these resultssuggest that 10 might coordinate to and stabilize both oxidized(as crystallized) and reduced clusters (the cluster is likelyreduced in the X-ray beam, as observed with 3),9 raising ofcourse the question: how does 10 bind to the fourth iron?While it is not possible to unequivocally assign the two ligandatoms that are close to the fourth Fe to C or N (since the ringcould be flipped), the resonance structures shown in Scheme 3seem likely since they permit η2 interactions with the cluster. Inaddition, as we shall describe below, the 180 degree ring-flippedstructures cannot account for the HYSCORE results.While it is possible that the aromatic ring in 10 might simply

be forming a van der Waals complex, the following observationssuggest that the pyridine ring is covalently bound to the cluster:(i) the short metal−ligand distances; (ii) the continuouselectron density; (iii) the stabilization of the fourth iron; and(iv) the higher inhibition activity with 10 as compared to 11−

13, reported previously with Aquifex aeolicus IspH.23 Based onour X-ray diffraction results we propose a η2 side-on metal−ligand coordination in the IspH−10 complex. This type ofbonding is seen with pyridine ligands coordinating to othermetals such as Ta,27,28 and similar bonding has also beenproposed to occur in Rh-catalyzed reactions.29

An interaction of the fourth iron with a nitrogen-containingligand has been observed previously with the substrate analogue4.20 Superposition of the IspH−10 and IspH−4 complexes(Figure S3 in the Supporting Information) shows that thepositions of the nitrogen atoms are similar, even though thecore structure of the molecules and their bonding showsignificant differences. Both compounds permit one watermolecule to occupy the remaining space in the active site, andthey are both able to stabilize the apical iron of the [Fe4S4]cluster, supporting the idea that the ability to interact with thefourth iron in the cluster is key for enhanced IspH inhibition.This intriguing X-ray structure raised several questions:

Which orientation does the pyridine ring adopt in themolecules with a [Fe4S4] cluster? What impact do substituentsat the pyridine ring have? Is the structure consistent withprevious spectroscopic observations? To further investigatethese questions we determined the structures of E. coli IspH incomplex with (pyridin-4-yl)methyl diphosphate (11), (pyridin-2-yl)methyl diphosphate (12), and (6-chloropyridin-3-yl)-methyl diphosphate (13). Compounds 11−13 are also pyridineanalogues carrying a methylene diphosphate substituent in the3- (10, 13), 4- (11), or 2- (12) position with 13 being the 6-Clanalogue of 10 (Scheme 2). The structures of the IspH−ligandcomplexes were solved to 1.8 Å resolution (11, PDB ID4MUY), 1.9 Å resolution (12, PDB ID 4MV0), and 1.9 Åresolution (13, PDB 1D 4MV5). All four species (10−13) bind

Figure 3. X-ray structure of IspH bound to (pyridin-3-yl)methyl diphosphate, 10. (a) A 2FO − FC electron density map (blue mesh, contoured at1.0σ) is presented for the [Fe4S4] cluster, the ligand, and the solvent molecules in the first coordination sphere. The protein is shown as cartoon andthe iron−sulfur cluster as a ball-and-stick model; the ligand and amino acid side chains are shown as stick models; the water molecule is shown as ared sphere, and the dotted lines indicate hydrogen or coordinative bonds. (b) A FO − FC electron density map (purple mesh, contoured at 3.0 σ)obtained by omitting the [Fe4S4] cluster and the ligand from the refinement. (c) Comparison between X-ray and docking results from ref 26.

Scheme 3. Proposed Binding Mode of the Pyridine Ring of10 to the [Fe4S4] Cluster in the IspH−10 Complex

Journal of the American Chemical Society Article

dx.doi.org/10.1021/ja501127j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXXC

in a similar manner to the SXN motif in the IspH active sitewith the diphosphate groups located in the same pocket as seenwith diverse diphosphates.Unlike the IspH−10 complex with a [Fe4S4] cluster, the

complexes of 11−13 (Figure 4) contain only a [Fe3S4] cluster.In a formal sense it is not known whether the [Fe3S4] cluster isselectively crystallized in the 11−13 complexes, or whether thefourth iron is simply lost during crystallization of the oxidizedprotein with the cluster being stabilized by binding to 10. Twopieces of information suggest that the latter possibility is morelikely. First, in four crystal structures of IspH−ligand complexesin which there is clear evidence for the presence of a fourthiron, the allyl group-containing ligands 3−59,12,20 and theenolate of 821 bind to (or very close to) the apical iron. Thissuggests that the fourth iron is stabilized by such ligand binding,and indeed with the complexes of 3 and the enolate of 8, the[Fe4S4] cluster is remarkably stable under aerobic conditions.However, the IspH products 1 and 2 crystallize as [Fe3S4]complexes, presumably due to weak product bonding(facilitating product release). Second, in previous work withAquifex aeolicus IspH we found that 10 is a better inhibitor thanare the remaining compounds investigated here.23 If 11−13caused cluster degradation, they would presumably be verypotent inhibitors, which they are not.23 We thus next considerthe structures of 11−13 in more detail.Based on all of the X-ray structures determined here it seems

unlikely that a para-N in a pyridine ring would bind to theapical Fe. The X-ray structure of IspH−11 indicates that thepyridine ring is positioned close to the [Fe3S4] cluster creatinga space for two water molecules. Comparison between thestructures of ligands 10 and 12 is intriguing since the nitrogenand the carbon atom that participate in the proposed η2

coordination in the IspH−10 structure are “switched” (Figure5). The difference electron-density (FO − FC) map and metalsite validation show, however, that the occupancy of the apicalsite is zero. Additionally, a B factor analysis suggests that thepyridine ring is in the same orientation as seen with 10 (bothnitrogen atoms point in the same direction). The refinementwith the aromatic ring flipped to the opposite site leads tosignificantly increased B factors, especially for the nitrogenatom. The preference of the ring for this orientation might be aresult of the increased hydrophobicity of one part of the activesite pocket with residues Val15, Val40, Ala73, and Val99 versusthe more hydrophilic site with residues Glu126, Thr167,

Thr168, Thr200, and Asn227. The 6-Cl analogue of 10, 13,would not be expected to have any significant σ-donor or H-bonding capacity due to the electron-withdrawing chlorineatom, expected to result in a decrease in pKa of ∼4.8 units forthe conjugate acid, and as expected the fourth iron of thecluster is absent in the IspH−13 complex structure.

Quantum Chemical Calculations. In order to investigatethe nature of the proposed iron−pyridine interaction in moredetail, we carried out a series of quantum mechanical (QM)calculations. It should be noted that both the redox state of thecluster and the protonation state of the ligand are formallyunknown, since oxidized clusters can be reduced by the X-raybeam, and protons are not detected in protein X-raycrystallography. Therefore, we investigated the four possiblecombinations of oxidized (Ox, S = 0) and reduced (Red, S = 1/2) iron−sulfur clusters with neutral pyridine (pyr) orpyridinium (pyrH+) ligands, including 180 degree pyridinering-flipped isomers. We also studied four different structuralmodels in order to see how the protein environment mightaffect bonding. In all cases the initial geometries were takenfrom the X-ray structures determined here, and we used thequantum chemical methods and program reported previouslyfor other [Fe4S4] systems.

23,30,31

At the so-called Small-Fopt level, the models consist ofoxidized or reduced [Fe4S4(SCH3)4]

n− (Small) clusters withthe protein environment being reduced to the cysteine sidechains truncated at the Cβ position. Both pyr and pyrH+

ligands were considered, in which all atoms were fullyoptimized (=F in Fopt) without any restraints. As shown in

Figure 4. Crystal structures of IspH in complex with (a) (pyridin-4-yl)methyl diphosphate, 11; (b) (pyridin-2-yl)methyl diphosphate, 12; (c) (6-chloropyridin-3-yl) methyl diphosphate), 13. Orientation and representation are according to Figure 3.

Figure 5. Structural superposition of IspH in complex with ligands10−12. (a) Front view of the pyridine ring and the side chains ofThr167 as well as E126. (b) Side view of the pyridine rings.

Journal of the American Chemical Society Article

dx.doi.org/10.1021/ja501127j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXXD

Table S3 in the Supporting Information, in these models, thefinal coordination modes in the optimized structures dependonly on the protonation state of the pyridine and not on theredox state of the cluster. These results show that, in suchunrestricted optimizations with no protein environment,pyridine prefers end-on coordination through the nitrogen’slone pair, as expected. The pyridinium on the other hand canstably adopt the side-on coordination, because of theunavailability of the nitrogen’s lone pair due to protonation.Interestingly, when the three cysteine Cβ atoms and the twocarbons neighboring the iron-bonded N and C atoms are fixedat the X-ray structure during geometry optimizations (calledSmall-Popt1, where P = Partial), the pyridine can now adoptside-on coordination, albeit only in the reduced state of the[Fe4S4] cluster (Table S3 in the Supporting Information).However, once the unique fourth iron is also fixed at the crystalstructure position (Small-Popt2 models in Table S3 in theSupporting Information), even with the oxidized cluster, theside-on mode is possible for pyridine. The average energypenalties due to fixing iron between Small-Popt1 and Small-Popt2 models are ca. 2 kcal/mol. These results suggest that theprotein environment may play a role in determining thecoordination mode. To probe this possibility in a more directway, we next investigated models containing Thr167, whichmay form a hydrogen bond to the nitrogen of the pyridinemoiety in 10. To better mimic the protein environment effect32

we also extended the residue models up to Cα for Thr167 andthe three coordinated cysteine residues, with only the Cα atomsbeing fixed at the X-ray structure positions. The pyridine wasalso modified to include the attached CH2OH group(mimicking the CH2OPP group), with only the oxygen fixedat the X-ray position. The results obtained using this structuralmodel are summarized in Table 1 and are displayed in Figure 6.For the oxidized cluster, only the pyridinium model is

consistent with the X-ray structure. However, if the cluster isreduced (in the X-ray beam, as reported earlier for 3),9 thenboth the pyridine and the pyridinium structures have Fe−Nand Fe−C distances close to the experimentally observedvalues. Based on a comparison with the previously reportedexperimental 14N isotropic hyperfine coupling (Aiso

N) andnuclear quadrupole coupling constant (NQCC = e2qQ/h)results for 10 bound to A. aeolicus IspH26, the pyridine model isclearly preferred over the pyridinium model. Both the predictedAiso

N and NQCCN values are in excellent agreement withexperiment, while the Aiso

N value for the pyridinium structure ispoorly predicted (Table 1). Specifically, the error in Aiso

N isonly ∼16% and that for the NQCC is ∼11%. Moreover,calculations based on the ring-flipped isomers (in which N is

not bonded to the fourth Fe) result in very small 14N hyperfinecouplings (∼0.2 MHz, to be compared with 6.6 MHz for thealternate conformer and 7.4 MHz from experiment). Theseresults all suggest that 10 binds to the reduced cluster in a side-on pyridine conformation, as found with other pyridinecomplexes.27,28

Overview of Ligand Binding to IspH. The structuresreported here are of general interest since they suggest anotherunusual type of metal−ligand bonding in IspH. In earlier workit was found that the substrate 3 binds to oxidized IspH via its1-OH group, as illustrated in Scheme 4. The same type of η1

bonding is seen in the amino and thiol analogues of 3 bound tooxidized IspH, as well as in the enolate intermediate that formsfrom the acetylene 6 on hydration. Higher (η2) coordinationnumbers are now apparent in the pyridine complex 10, and ithas also been proposed that η2 and η3 species14 are involved asreactive intermediates in IspH catalysis, based on the results ofHYSCORE and electron−nuclear double resonance (ENDOR)spectroscopies and of DFT calculations.14,33 Plus, a putativeferraoxetane reactive intermediate (containing Fe−C and Fe−O bonds) in the IspG reaction15 (in which methyl-erythritolcyclo-diphosphate is converted to 3) has been proposed, againbased on the results of HYSCORE and ENDOR spectroscopiesand DFT calculations.34 These structures are in many cases ofinterest from a mechanism of action/inhibition perspective, andvia protein engineering, they may lead to new chemistries.

Table 1. Geometric and Spectroscopic Properties of the Structural Models Shown in Figure 6

coord mode RFeN/C (Å) RFeC (Å) AisoN (MHz) NQCCN (MHz)

expta η2 2.3 2.4exptb reduced 7.4 3.0calc 1 oxidized pyr η1 2.1 3.0 3.0

2 oxidized pyrH+ η2 2.1 2.1 1.73 reduced pyr η2 2.0 2.2 6.6 3.54 reduced pyrH+ η2 2.1 2.0 1.2 2.45 oxidized pyr (flipped) η1 2.3 2.8 5.06 oxidized pyrH+ (flipped) η2 2.1 2.1 2.67 reduced pyr (flipped) η2 2.2 2.2 0.2 4.58 reduced pyrH+ (flipped) η2 2.1 2.1 0.5 3.3

aThis work. bReference 26.

Figure 6. Optimized structures for oxidized and reduced models withpyridine or pyridinium ligands. Color scheme: C, cyan; Fe, black; S,yellow; O, red; H, gray.

Journal of the American Chemical Society Article

dx.doi.org/10.1021/ja501127j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXXE

■ CONCLUSIONSWe report the first crystal structures of E. coli IspH bound topyridine ligands. Of the four structures obtained, only IspHbound to 10 contains a [Fe4S4] cluster (the others contain[Fe3S4] clusters) and 10 was previously reported to be the mostpotent inhibitor of A. aeolicus IspH.23 Two atoms in the ligand’spyridine ring bind close (2.3−2.4 Å) to the fourth iron in thecluster, and the general pose is similar to that predicted earlierbased on docking calculations (a 0.89 Å rmsd for all ligand non-hydrogen atoms). QM results enable very good predictions ofprevious 14N hyperfine coupling (Aiso

N = 7.4 MHz, experiment;6.6 MHz, calculated) as well as the 14N quadrupole couplingconstant results (NQCC = 3.0 MHz experiment; 3.5 MHz,calculated) from HYSCORE spectroscopy, and strongly suggestthat the [Fe4S4] cluster is in a reduced state with the pyridineinvolved in side-on η2 coordination. Overall, the resultscontribute to the growing database of bio-organometalliccomplexes of IspH, of interest not only from a drug discoveryperspective but potentially from a synthetic chemistryperspective as well.

■ EXPERIMENTAL SECTIONProtein purification. E. coli IspH containing a His6 tag (encoded

in the pQE30 plasmid) was coexpressed with isc proteins (encoded inthe pACYC184 plasmid) in XL-1 blue cells. Cells were grown at 30 °Cfor 20 h in Terrific Broth supplemented with ampicillin (180 mg/L),chloramphenicol (25 mg/L), L-cysteine (167 mg/L), and ammoniumiron(III) citrate (32 mg/L). Cells were then harvested bycentrifugation, washed with 0.9% aqueous sodium chloride, andstored at −80 °C. All further steps were performed under anaerobicconditions in a Coy Vinyl Anaerobic Chamber under an N2/H2 (95%/5%) atmosphere. The cell pellet was resuspended in five volumes of100 mM Tris hydrochloride, pH 8, containing 500 mM sodiumchloride, 20 mM imidazole, and 5 mM sodium dithionite. The mixturewas passed through a French Press and was subsequently centrifuged.The resulting crude extract was loaded onto a Ni2+ ChelatingSepharose Fast Flow column (20 mL), which had been equilibrated in100 mM Tris hydrochloride, pH 8, containing 500 mM NaCl and 20mM imidazole. The protein was eluted with 100 mM Trishydrochloride, pH 8, containing 500 mM NaCl and 150 mMimidazole. The solution was dialyzed against 50 mM Tris hydro-chloride, pH 8.Crystallization. E. coli IspH (18.3 mg/mL) was incubated with an

aqueous solution of the ligand (final concentration 1 mM) for 20 minprior to setting up crystal trays. Brown crystals were obtained by using

the sitting-drop vapor diffusion method with a 1:1 ratio of protein andreservoir solution at 20 °C. The precipitant consisted of 100 mMBisTris/HCl, pH 6.5, 200 mM ammonium sulfate, and 25%polyethylene glycol 3350. Crystals were soaked with cryoprotectantsolution (50% aqueous polyethylene glycol 400) for 1 min, mountedon loops, and flash frozen in a stream of nitrogen gas at 100 K (OxfordCryo Systems).

Data Collection and Structure Determination. Native data setswere collected using synchrotron radiation at the X06SA-beamline atthe Swiss Light Source, Villigen, Switzerland. The coordinates of E. coliIspH bound to the substrate HMBPP (PDB ID 3KE8) were used forphasing by molecular replacement.16,35 Data sets were processed usingthe program package XDS,36 and anisotropy of diffraction wascorrected using TLS37,38 refinement. Model building and refinementwere performed with Coot39 and Refmac.38,40 The occupancy of theapical iron site was refined using phenix.refine41 and validated withCheckMyMetal web server.42 Electron density maps were calculatedusing FFT,38,43 Ramachandran plots were calculated with PRO-CHECK,38,44 and figures were prepared using PyMOL.45 For moredetails see Table S1 in the Supporting Information.

Accession Numbers. The atomic coordinates for IspH bound tothe ligands 10, 11, 12, and 13 have been deposited in the Protein DataBank, Research Collaboratory for Structural Bioinformatics at RutgersUniversity (IspH−10 PDB ID 4MUX, IspH−11 PDB ID 4MUY,IspH−12 PDB ID 4MV0, and IspH−13 PDB ID 4MV5).

Computational Details. To investigate the binding of 10 to theiron−sulfur cluster, we used density functional theory. All calculationswere performed with the Gaussian 09 program46 using a Wachters’basis for Fe,47 6-311G(d) for other heavy atoms, 6-31G(d) forhydrogens, with the BPW9148,49 functional, as reported previ-ously.23,30,31,49 This approach23,50−56 has enabled accurate predictionsof Moßbauer and EPR/nuclear magnetic resonance (NMR) hyperfineproperties in a diverse range of iron-containing proteins and modelsystems. The predicted isotropic EPR hyperfine coupling and nuclearquadrupole coupling constant results (Table S3 in the SupportingInformation) are from the directly calculated values from the quantumchemical calculations, which were scaled based on the regression linesbetween theory and experiment, as described in previous reports.26,31

The coordinates of all of the optimized structures described in the textare provided in Tables S4−S23 in the Supporting Information.

■ ASSOCIATED CONTENT*S Supporting InformationAdditional tables with data collection and refinement statistics;geometric and spectroscopic properties of various QM models;coordinates of all quantum chemically optimized models.Supplementary figures and references. This material is availablefree of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Authorspan.ingrid@gmail.comPresent Address‡Department of Molecular Biosciences, Northwestern Univer-sity, 2205 Tech Drive, Evanston, Illinois 60208, United States.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by the TUM Graduate School, theHans-Fischer Gesellschaft, DFG Grant GR1861/5-1, theUnited States Public Health Service (NIH Grants GM085774to Y.Z. and GM065307 to E.O.), and National ScienceFoundation Grant CHE-1300912 to Y.Z.. We would like tothank the staff of the X06SA-beamline at the Paul SchererInstitute (Swiss Light Source in Villigen, Switzerland) for

Scheme 4. Overview of the Observed or Proposed BindingModes of Various Ligands to the Unique Fourth Iron of the[Fe4S4] Cluster of IspH (or for 15, IspG)a

aThe binding modes can be subdivided into (a) η1-, (b) η2-, and (c),η3-coordination.

Journal of the American Chemical Society Article

dx.doi.org/10.1021/ja501127j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXXF

support during data collection, as well as Professor Jared C.Lewis for helpful comments.

■ REFERENCES(1) Ruzicka, L. Cell. Mol. Life Sci. 1953, 9, 357.(2) Bloch, K. Steroids 1992, 57, 378.(3) Rohmer, M.; Knani, M.; Simonin, P.; Sutter, B.; Sahm, H.Biochem. J. 1993, 295 (Pt 2), 517.(4) Rohdich, F.; Zepeck, F.; Adam, P.; Hecht, S.; Kaiser, J.; Laupitz,R.; Grawert, T.; Amslinger, S.; Eisenreich, W.; Bacher, A.; Arigoni, D.Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 1586.(5) Adam, P.; Hecht, S.; Eisenreich, W.; Kaiser, J.; Grawert, T.;Arigoni, D.; Bacher, A.; Rohdich, F. Proc. Natl. Acad. Sci. U.S.A. 2002,99, 12108.(6) Altincicek, B.; Duin, E. C.; Reichenberg, A.; Hedderich, R.;Kollas, A. K.; Hintz, M.; Wagner, S.; Wiesner, J.; Beck, E.; Jomaa, H.FEBS Lett. 2002, 532, 437.(7) Xiao, Y.; Liu, P. Angew. Chem., Int. Ed. 2008, 47, 9722.(8) Xiao, Y.; Zhao, Z. K.; Liu, P. J. Am. Chem. Soc. 2008, 130, 2164.(9) Grawert, T.; Span, I.; Eisenreich, W.; Rohdich, F.; Eppinger, J.;Bacher, A.; Groll, M. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 1077.(10) Wang, W.; Wang, K.; Liu, Y. L.; No, J. H.; Li, J.; Nilges, M. J.;Oldfield, E. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 4522.(11) Xiao, Y.; Chang, W. C.; Liu, H. W.; Liu, P. Org. Lett. 2011, 13,5912.(12) Span, I.; Grawert, T.; Bacher, A.; Eisenreich, W.; Groll, M. J.Mol. Biol. 2012, 416, 1.(13) Citron, C. A.; Brock, N. L.; Rabe, P.; Dickschat, J. S. Angew.Chem., Int. Ed. 2012, 51, 4053.(14) Xu, W.; Lees, N. S.; Hall, D.; Welideniya, D.; Hoffman, B. M.;Duin, E. C. Biochemistry 2012, 51, 4835.(15) Rekittke, I.; Wiesner, J.; Rohrich, R.; Demmer, U.; Warkentin,E.; Xu, W.; Troschke, K.; Hintz, M.; No, J. H.; Duin, E. C.; Oldfield,E.; Jomaa, H.; Ermler, U. J. Am. Chem. Soc. 2008, 130, 17206.(16) Grawert, T.; Rohdich, F.; Span, I.; Bacher, A.; Eisenreich, W.;Eppinger, J.; Groll, M. Angew. Chem., Int. Ed. 2009, 48, 5756.(17) Kent, T. A.; Dreyer, J. L.; Kennedy, M. C.; Huynh, B. H.;Emptage, M. H.; Beinert, H.; Munck, E. Proc. Natl. Acad. Sci. U.S.A.1982, 79, 1096.(18) Layer, G.; Heinz, D. W.; Jahn, D.; Schubert, W.-D. Curr. Opin.Chem. Biol. 2004, 8, 468.(19) Tamarit, J.; Gerez, C.; Meier, C.; Mulliez, E.; Trautwein, A.;Fontecave, M. J. Biol. Chem. 2000, 275, 15669.(20) Span, I.; Wang, K.; Wang, W.; Jauch, J.; Eisenreich, W.; Bacher,A.; Oldfield, E.; Groll, M. Angew. Chem., Int. Ed. 2013, 52, 2118.(21) Span, I.; Wang, K.; Wang, W.; Zhang, Y.; Bacher, A.; Eisenreich,W.; Li, K.; Schulz, C.; Oldfield, E.; Groll, M. Nat. Commun. 2012, 3,1042.(22) Oldfield, E. Acc. Chem. Res. 2010, 43, 1216.(23) Wang, K.; Wang, W.; No, J. H.; Zhang, Y.; Oldfield, E. J. Am.Chem. Soc. 2010, 132, 6719.(24) Janthawornpong, K.; Krasutsky, S.; Chaignon, P.; Rohmer, M.;Poulter, C. D.; Seemann, M. J. Am. Chem. Soc. 2013, 135, 1816.(25) Ahrens-Botzong, A.; Janthawornpong, K.; Wolny, J. A.;Tambou, E. N.; Rohmer, M.; Krasutsky, S.; Poulter, C. D.;Schunemann, V.; Seemann, M. Angew. Chem., Int. Ed. 2011, 50, 11976.(26) Wang, W.; Li, J.; Wang, K.; Smirnova, T. I.; Oldfield, E. J. Am.Chem. Soc. 2011, 133, 6525.(27) Covert, K. J.; Neithamer, D. R.; Zonnevylle, M. C.; LaPointe, R.E.; Schaller, C. P.; Wolczanski, P. T. Inorg. Chem. 1991, 30, 2494.(28) Neithamer, D. R.; Parkanyi, L.; Mitchell, J. F.; Wolczanski, P. T.J. Am. Chem. Soc. 1988, 110, 4421.(29) Wiedemann, S. H.; Lewis, J. C.; Ellman, J. A.; Bergman, R. G. J.Am. Chem. Soc. 2006, 128, 2452.(30) Liu, Y. L.; Guerra, F.; Wang, K.; Wang, W.; Li, J.; Huang, C.;Zhu, W.; Houlihan, K.; Li, Z.; Zhang, Y.; Nair, S. K.; Oldfield, E. Proc.Natl. Acad. Sci. U.S.A. 2012, 109, 8558.(31) Li, J.; Wang, K.; Smirnova, T. I.; Khade, R. L.; Zhang, Y.;Oldfield, E. Angew. Chem., Int. Ed. 2013, 52, 6522.

(32) Yang, L.; Ling, Y.; Zhang, Y. J. Am. Chem. Soc. 2011, 133, 13814.(33) Wang, W.; Wang, K.; Span, I.; Jauch, J.; Bacher, A.; Groll, M.;Oldfield, E. J. Am. Chem. Soc. 2012, 134, 11225.(34) Wang, W.; Li, J.; Wang, K.; Huang, C.; Zhang, Y.; Oldfield, E.Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 11189.(35) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J.; Meyer, E. F., Jr.;Brice, M. D.; Rodgers, J. R.; Kennard, O.; Shimanouchi, T.; Tasumi,M. Eur. J. Biochem. 1977, 80, 319.(36) Kabsch, W. J. Appl. Crystallogr. 1993, 26, 795.(37) Winn, M. D.; Isupov, M. N.; Murshudov, G. N. Acta Crystallogr.D 2001, 57, 122.(38) Winn, M. D.; Ballard, C. C.; Cowtan, K. D.; Dodson, E. J.;Emsley, P.; Evans, P. R.; Keegan, R. M.; Krissinel, E. B.; Leslie, A. G.;McCoy, A.; McNicholas, S. J.; Murshudov, G. N.; Pannu, N. S.;Potterton, E. A.; Powell, H. R.; Read, R. J.; Vagin, A.; Wilson, K. S.Acta Crystallogr. D 2011, 67, 235.(39) Emsley, P.; Cowtan, K. Acta Crystallogr. D 2004, 60, 2126.(40) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta Crystallogr.D 1997, 53, 240.(41) Afonine, P. V.; Grosse-Kunstleve, R. W.; Echols, N.; Headd, J. J.;Moriatry, N. W.; Mustyakimov, M.; Terwilliger, T. C.; Urzhumtsev, A.;Zwart, P. H.; Adams, P. D. Acta Crystallogr. D 2012, 68, 352.(42) Zheng, H.; Chordia, M. D.; Cooper, D. R.; Chruszcz, M.;Muller, P.; Sheldrick, G. M.; Minor, W. Nat. Protoc. 2014, 9, 156.(43) Read, R. J.; Schierbeek, A. J. J. Appl. Crystallogr. 1988, 21, 490.(44) Laskowski, R. A.; Rullmannn, J. A.; MacArthur, M. W.; Kaptein,R.; Thornton, J. M. J. Biomol. NMR 1996, 8, 477.(45) The PyMOL Molecular Graphics System, Version 1.3; Schrodinger,LLC.(46) Frisch, M. J. et al. Gaussian 09, Revision B.01; Gaussian, Inc.:Wallingford, CT, 2010.(47) http://www.emsl.pnl.gov/forms/basisform.html.(48) Becke, A. D. Phys. Rev. A 1988, 38, 3098.(49) Perdew, J. P.; Burke, K.; Wang, Y. Phys. Rev. B 1996, 54, 16533.(50) Zhang, Y.; Oldfield, E. J. Am. Chem. Soc. 2004, 126, 4470.(51) Mao, J.; Zhang, Y.; Oldfield, E. J. Am. Chem. Soc. 2002, 124,13911.(52) Zhang, Y.; Mao, J.; Godbout, N.; Oldfield, E. J. Am. Chem. Soc.2002, 124, 13921.(53) Zhang, Y.; Mao, J.; Oldfield, E. J. Am. Chem. Soc. 2002, 124,7829.(54) Zhang, Y.; Gossman, W.; Oldfield, E. J. Am. Chem. Soc. 2003,125, 16387.(55) Zhang, Y.; Oldfield, E. J. Phys. Chem. A 2003, 107, 4147.(56) Zhang, Y.; Oldfield, E. J. Phys. Chem. B 2003, 107, 7180.

Journal of the American Chemical Society Article

dx.doi.org/10.1021/ja501127j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXXG